Output Node

The Output node is where image generation actually happens. It receives a prompt and (optionally) one or more source images, runs them through the model you’ve selected, and shows the result.

Targets

• prompt — connect from a Prompt node
• images — connect one or more Image / Output / Video nodes

If you only connect a prompt → text-to-image. If you connect both → image-to-image editing or multi-image fusion (model-dependent).

Picking a model

Each Output node has its own model dropdown. Switching models doesn’t replace the result — it just changes what the next click of Generate will produce.

Common picks:

• Nano Banana 2 — fast, cheap, supports T2I and I2I (8 credits)
• Nano Banana Pro — higher fidelity, longer prompt context
• GPT Image 2 — OpenAI; strong prompt adherence, ultra-wide aspect ratios (5 credits)
• Imagen 4 Ultra — Google’s flagship image model

See Image Models for the full lineup and per-model parameters.

Generating

Click Generate. Banana Flow will:

1. Verify you have enough credits
2. Send the request to the model provider
3. Poll until the image is ready (typically 5–30 seconds)
4. Store the result in your Library and display it on the node

Editing the result

Click Edit on the generated image to open the in-canvas editor. Saving spawns a new Image node — your original Output result is preserved.

Bulk generation

There’s no special mode to enable. The moment your connected prompt contains a {a|b|c} wildcard, the next click of Generate is auto-routed to the bulk pipeline. The Output node shows a chip below the prompt with the live combo count and total cost. See the Bulk Generation tutorial.